Spinning Raw Data Into Synthetic (Revenue) Gold For FIs

Data is everywhere, there for the taking, turning insight into top-line growth. For financial institutions (FIs) and other FinTech companies, monetizing that data effectively and, above all, safely can be a tightrope walk — balancing risk and reward.

Done correctly, the same sets of consumer data, de-risked, can yield revenue opportunities that come from both outside sources — as those firms sell that data to third parties — and within the firms, as executives spend less time on compliance activities and more time on innovation.

In an interview with PYMNTS, Randy Koch, CEO of ARM Insight, said a successful data monetization strategy can take shape along four key principles.

First, executives must strive to comply with security and regulation mandates, which dictate how data is collected and ultimately used. Second, executives must aggregate and transform data. Third, the most important principle, he noted, is “understanding that there are three types of data: raw data, anonymous data and synthetic data.” Last, “you have to know how to monetize that data, and what your monetization options are,” he said.

In that quest to monetize data, there are two paths that can be pursued, he explained. On one hand, there’s the goal of monetizing data externally, which is selling the data to third parties. On the other hand, there is the ability for a firm — such as an FI or processor — to monetize data internally.

The Process And Types Of Data

Regardless of whether banks and other firms are pursuing internal or external monetization, Koch said it is important for companies to “clean up” their data, get rid of duplicates and place that data into a data warehouse, where it can be made available for modification. The next and more important step is to create raw, anonymous and synthetic data, “because there is a big security, compliance and privacy impact.”

Delving into the data segmentation, ARM Insight has classified raw data as among the most valuable types of data, but also the most risky. It typically includes credit card data, addresses, names and even Social Security numbers. Anonymized data has had the aforementioned personal information removed, but still has some descriptive information about transactions.

Synthetic data, as previously described in this space, consists of data sets that take anonymized raw data and run through tech-driven mechanisms, such as machine learning (ML) or artificial intelligence (AI). New data sets are created that cannot be traced back to consumers, the places where they do business or even the FI tied to the transaction. Koch explained that synthetic data is a fake data set that “doesn’t have any risk to the raw data set, yet it retains data accuracy of the original data set.”

With those three data types in hand, he said, “Ninety percent of what you can monetize, and actually reduce risk on, is the synthetic data. That’s where your focus should be, because it has the least amount of risk and the most amount of upside for monetization.”

The External Route

Though the external monetization option has been growing, Koch noted that executives still fear using data (especially raw data). In the past five to 10 years, using consumer information has been viewed as a risk and liability. That fear has only been exacerbated by a seemingly endless spate of headlines surrounding data breaches, and the looming specter of fines and reputational risks.

“When you talk about data monetization,” he said, “everyone thinks of raw data,” and the risk inherent with that data.

Those fears, he added, can be solved through the use of synthetic data. In short, concerns with GLBA, GDPR, PCI, reputational and headline risks tied to raw data can literally “all go away with the use of synthetic data.”

The remaining compliance issues lie with, what he termed, the internal contract that exists between the FI and its end customers. For example, he pointed to the contracts that exist between banks and customers, as the latter use debit and credit cards — where FIs cannot resell personal identification to any third party.

However, as long as data is anonymized, aggregated and depersonalized, it can be used for third-party purposes. Synthetic data, which sports all those features, can be created by FIs and used by retailers to, say, see how effectively they are targeting millennials.

Monetizing Data Internally

Synthetic data eliminates the single biggest risk housed within an FI: employee data misuse. As he noted, FIs may have hundreds of analysts and engineers looking at the firms’ core data sets, and employee misuse or abuse can pose problems. If the employees are working with synthetic data, he said, “you have reduced risk and increased security on the data itself.”

Safely housed within the firm, data can be seen as both asset and profit center, aiding in efforts to innovate. “The roadmap looks pretty straightforward,” said Koch, noting that improved analytics through ML and AI can help FIs provide customers with services and products that allow them to differentiate themselves from the competition. Getting over the aforementioned fears — and realizing that, with data, “10 percent of it can be scary (that raw data), but the other 90 percent of it is ready to be monetized” — can help executives see that monetization lies beyond simply selling to third parties.

“Once you understand those three data types, two things happen. You have increased security and increased compliance, because you’re working mainly with synthetic data and you have increased revenue streams. It really comes down to those simple points,” he explained.

Both the internal and external routes “are growing at a good clip now,” he added. “We are still at the beginning stages of the long-haul data monetization journey.”